ABSTRACT Project 3 Enhancing crystallization using natural binding partners Crystallization is the major bottleneck in structure determination by X-ray crystallography. If we could overcome this bottleneck, we could immediately increase the number of structures solved by nearly 3-fold. This proposal seeks to improve the crystallization of proteins and protein complexes by finding and supplying missing partner proteins. A novel and important concept is the two-tiered approach of bioinformatics prediction followed by a high-throughput experimental validation of protein complex formation. A second unique core approach is the use of experimental high-throughput split GFP cell-based assays to directly identify interaction proteins. Validation is a vital component, and will allow an accurate determination of the relative success rate of the different core approaches. The experimental high-throughput identification of binding partners will be developed using the E. coli genome since it (1) has many known protein complexes, allowing us to validate the approach, and (2) is predicted to have nearly 550 failed targets in the PSI Large Scale Centers that failed because they are missing partners. We will develop general procedures for identification of natural binding partners of a target protein using bioinformatics. An inference engine will be used to generate lists of potential binding partners and an associated probability or confidence estimator. In parallel, we will develop a highthroughput library-based procedure for identification of natural binding partners using in-vivo reporters. We will validate the experimental procedure using protein-protein interaction assays on control complexes and apply it to find missing partners from the E. coli genome for E. coli targets that are likely to have failed in PSI centers due to missing proteins. We will further develop a procedure for identification or validation of members of natural complexes with a pull-down assay based on split GFP. We will validate all these procedures by using them to identify, stabilize, and determine structures of targets of structural genomics Large-Scale Centers that have failed at the stage of crystallization. The approaches developed in this work will lead to methods that will increase our understanding of human health and our ability to cure human illness by facilitating the structural determination of proteins involved in human disease.